Chapter 5: Complementary Models

Myths, Models and Paradigms: A Comparative Study in Science and Religion
by Ian Barbour

Chapter 5: Complementary Models

In he preceding chapters the functions and the status of models in science and in religion were discussed. I now wish to look more specifically at the role of complementary models in twentieth century physics, and then at some possible parallels in religious thought. Can one continue to employ two very different models within either science or religion? Can an electron be thought of as both a wave and a particle? Can one use both personal and impersonal models of Ultimate Reality?

1. The Wave-Particle Duality

We have seen that particle models, such as the billard-ball model, dominated the classical physics of matter. By the nineteenth century, another basic type of model, that of waves in continuous media, was also being employed for a different group of phenomena involving light and electromagnetism. But early in the present century a number of puzzling experiments seemed to call for the use of both wave and particle models for both types of phenomena. On the one hand, Einstein’s equations for the photo-electric effect and Compton’s work on photon scattering showed that light travels in discrete packets, with definite energy and momentum, behaving very much like particles. Conversely, electrons, which had always been viewed as particles, showed the spread-out interference effects characteristic of waves. Waves are continuous, extended, and interact in terms of phase; particles are discontinuous, localized, and interact in terms of momentum. There seems to be no way to combine them into one unified model.

Suppose, for example, that a beam of electrons is shot through two narrow slits in a metal screen and strikes a photographic film placed a few centimeters behind the screen. Each electron registers as a single tiny dot on the film; it seems to arrive as a particle, and it must presumably have gone through either one slit or the other if the charge and mass of the electron are indivisible. Yet the dots on the film fall in an interference pattern of parallel bands, which can be explained only if one assumes a wave passing through both slits.1 This same wave-particle duality is found throughout atomic physics. ‘Compromise models’, such as localized wave packets, provide no general resolution of the paradox. But a unified mathematical formalism can be developed which allows the observed events to be predicted statistically. One can calculate the probability that an electron will strike the film at any given point. Within the calculated probability-distribution, however, the point at which a particular electron will strike is not predictable at all.

Similarly, no unified model of the quantum atom has been developed. The earlier ‘Bohr model’ of the atom could be easily visualized; particle-like electrons were thought to follow orbits around the nucleus, resembling a miniature solar system. But the atom of quantum mechanics is not picturable at all. One might try to imagine patterns of probability waves filling the space around the nucleus like some three-dimensional symphony of musical tones of incredible complexity, but the analogy would not help us much. The atom is not just inaccessible to direct observation and unimaginable in terms of sensory qualities; it cannot even be described coherently in terms of classical concepts such as space, time and causality. The domain of the very small must be radically different from the domain of everyday objects. We can describe statistically by differential equations what happens in experiments, but we cannot ascribe familiar attributes consistently to the inhabitants of the atomic world. We seem to be a long way from the familiar ‘billiard ball’ model of nineteenth-century physics.

In quantum physics, then, models are only remotely and inconsistently related to theories; and theories in turn have extremely indirect and in general only probabilistic connections with observations. In his work, the physicist relies on the unified mathematical formalism of the theory and abandons dependence on pictorial representations. He may be unhappy that the equations of his theory do not yield exact predictions about individual events; his probability-distributions reflect an indeterminacy in the simultaneous measurement of pairs of variables (such as position and momentum). But no one has devised any better alternative, and the physicist has learned to live with the theory he has. He talks about ‘states’ and ‘operators’ instead of particles or waves.

It is not surprising that the positivist finds in quantum physics support for his conviction that we should discard all model; and treat theories as mere calculational devices for correlating observations. He urges us to give up trying to imagine what goes on between observations; in the two-slit experiment, he says, it is useless to ask what the electron was doing before it hit the film. The highly abstract mathematical formalism should be treated as a mental construct for making predictions. The equations will allow us to correlate statistically the final and initial states of specified experimental situations, which are the only things we can observe. Models are superfluous and theories are useful fictions; neither is a representation of the world.

I would myself see modern physics as a strong warning against literalism rather than as evidence for the absolute rejection of models. Even the apparently bare formalisms of quantum theory are not ‘totally uninterpreted’, for they still carry imaginative associations. These may not be pictorial in character, but they do convey analogies with other fields which are important in suggesting rules of correspondence with observable variables. Nagel states:

There is of course no question whatever that the terminology of ‘particles’ and ‘waves’ is suggestive and heuristically valuable. Nevertheless, the usefulness of this terminology must not hide from us the fact that it is employed analogically and is not to be construed literally... Nevertheless, there are great psychological advantages in having such models for a theory. In consequence, with such models as their objective, physicists frequently formulate the content of quantum mechanics in the language of classically conceived particles and waves, because of certain analogies between the formal structures of classical and quantum mechanics... Accordingly, although a satisfactory uniformly complete interpretation of quantum mechanics based on a single model cannot be given, the theory can be satisfactorily interpreted for each concrete experimental situation to which the theory is applied.2

The analogies in quantum physics contribute to the interpretation of the formalism, its extension to new domains, and its possible modification. The Schrbdinger ‘wave equation’ cannot but remind one of the familiar second-order differential equations which apply to waves. ‘When the physical model of wave-motion in a material medium had to be abandoned in physics’, writes Mary Hesse, ‘it left its traces in the kind of mathematics which was used, for this was still a mathematical language derived from the wave equations of fluid motion, and so, for the mathematician, it carried some of the imaginative associations of the original physical picture.’3 Dirac’s theory, she maintains, presupposes a particle analogy ‘which invests the formalism with experiential relevance’; the assumption of ‘holes’ in postulated negative-energy states led him to predict the existence of the positron. Hutten compares such highly attenuated analogies with the lingering grin of the Cheshire cat: ‘The picture of the cat has receded into the background, but, knowing that there once was a cat, we understand that the residual phenomenon may be interpreted as a grin.’4 Even a few of the associations of waves and particles may provide useful clues concerning rules of correspondence.

Ivlesse notes further that the wave model’s positive analogy is the particle model’s negative analogy, and vice versa. Hence they do not conflict directly with each other -- even though they cannot be combined into one model because there is little overlap in the positive analogies. Yet if we retained only the mathematical formalism which is derived from the two positive analogies, we would lose the two neutral analogies which may provide promising ideas for further exploration. Thus implicit wave and particle models are not just psychological aids for the non-mathematical layman; they are of value to the scientist for extension of the theory, postulation of new correspondence rules, and applications to new types of observation. Hesse concludes:

If we were forbidden to talk in terms of models at all, we should have no expectations at all, and we should be imprisoned for ever inside the range of our existing experiments.... And it is in arguing in terms of these features that the particle and wave models are still essential, supplemented by the hunches that physicists have acquired about when to argue in terms of one and when the other. The particle and wave models themselves cannot he regarded as simply descriptive of reality, but when taken together in this complicated way they can be regarded as giving us knowledge of the real world.5

2. The Complementarity Principle

Niels Bohr has given the most influential defense for the retention of both wave and particle models along with the recognition of their limitations. ‘A complete elucidation of one and the same object’, he writes, ‘may require diverse points of view which defy a unique description.’6 There are three interwoven themes in his complex discussion of complementarity. First, he shows that the more a particular experimental arrangement makes wave-like behaviour evident, the less evident is particle-like behaviour, and vice versa. The extreme cases of unambiguous wave and particle behaviour occur in mutually exclusive laboratory situations.7 As one physicist puts it, you may have to use a wave model on Mondays, Wednesdays and Fridays, and a particle model on Tuesdays, Thursdays, and Saturdays. But I would point out that on some days (perhaps Sundays) you may have to use both models. The extreme cases are indeed mutually exclusive, but in a middle range features of both wave and particie seem to be manifest together in a single experiment, such as the two-slit-case above; neither model is adequate, yet neither can be dispensed with. The problem goes beyond that of using different models in different experiments.

A second theme frequently mentioned by Bohr is the interaction between subject and object in every experiment. He states that no sharp line can be drawn between the process of observation and what is observed. We are actors and not merely spectators, and we choose the experimental tools we will employ. The measuring procedure disturbs the system to be measured. But I would still ask: can this ‘influence of the observer’ account for the unpredictability of observations when nothing is done to disturb the system (e.g. spontaneous nuclear disintegration, or the ‘diffusion of a wave packet’)? The uncertainty relationship can in fact be derived from quantum theory without any reference to ‘disturbing the system’. Most of Bohr’s followers no longer claim that this second thesis covers all instances of indeterminacy. In any case it would not help us understand why complementary models are needed.

On this question a third theme in Bohr’s writing is more illuminating, namely his discussion of conceptual limitations in human understanding. Man as knower, rather than man as experimenter, is the centre of attention here. Bohr shares Kant’s scepticism about the possibility of knowing the world in itself. He holds that classical concepts are ‘forms of perception’ imposed by man. If we try, as it were, to force nature into certain conceptual moulds, we preclude the full use of other moulds. Thus we must choose between complete causal or spatio-temporal descriptions, between adequate wave or particle models, between accurate knowledge of position or momentum. The more one set of concepts is used, the less can the complementary set be applied simultaneously. We have successive and incomplete complementary pictures that cannot be neatly unified. This reciprocal limitation occurs because the atomic world cannot be described in terms of classical concepts, which, according to Bohr, are the only ones available to us. I would accept this stress on the importance of the categories of understanding imposed by the knower, but I would want to attribute them less to the given structures of the mind (as in Bohr’s neo-Kantian view) than to the limitations of our experience and imagination. We do not need to assume, with Bohr, that waves and particles exhaust the possible types of basic model.

Bohr himself proposes that the idea of complementarity could be extended to other phenomena susceptible to analysis by two kinds of model: mechanistic and organic models in biology, behaviouristic and introspective models in psychology, models of free will and determinism in philosophy, or of divine justice and divine love in theology.8 Some authors go further and speak of the complementarity of science and religion. Thus C. A. Coulson, after explaining the wave-particle duality and Bohr’s generalization of it, calls science and religion ‘complementary accounts of one reality’.9

D. M. MacKay has defended such extended user of the idea of complementarity. He defines two descriptions as complementary if (a) they have a common referent, (b) the logical preconditions for their use are mutually exclusive, and (c) each is in principle exhaustive in its own frame of reference. The wave-particle duality in physics is for him an instance of a more general logical relationship between two accounts of one object under differing conditions or from different perspectives. According to his definition, vertical and horizontal plans of a building are ‘complementary descriptions’. So are mental and physical descriptions of a person’s activity. MacKay concludes that science and religion may likewise be considered complementary.10

I am somewhat dubious about such extended usage of the term if it is intended to convey some parallel with complementarity in physics. I would want to set down several conditions for applying the concept of complementarity:

i. Complementarity provides no justification for an uncritical acceptance of dichotomies. It cannot be used to avoid dealing with inconsistencies or to veto the search for unity. The ‘paradoxical’ element in the wave-particle duality should not be over-emphasized. We do not say that an electron is both a wave and a particle, but that it exhibits wave-like and particle-like behaviour; moreover we do have a unified mathematical formalism which provides at least probabilistic predictions. And as Feyerabend insists, we cannot rule Out in advance the search for new unifying models (such as David Bohm’s postulation of sub-atomic causal mechanisms), even though previous attempts have not yielded any new theories in better agreement with the data than quantum theory.11 Coherence remains an important ideal and criterion in all reflective enquiry.

2. Models should be called complementary only if they refer to the same entity and are of the same logical type. Wave and particle are models of a single entity (e.g. an electron) in a single situation (e.g. a two-slit experiment); they are on the same logical level and had previously been employed in the same discipline. As Peter Alexander has pointed out, these conditions do not apply to ‘science and religion’.12 They do not refer to the same entity. They arise typically in differing situations and serve differing functions in human life. For these reasons I will speak of science and religion as alternative languages using alternative models, and restrict the term ‘complementary’ to models of the same logical type within a given language.

3. The complementarity of models, under these conditions, underscores the inadequacy of literalism. The use of one model limits the use of the other; they are not simply ‘alternative models’ having different domains or functions. They are symbolic representations of aspects of reality which cannot be consistently visualized in terms of analogies with everyday experience; they are only very indirectly related to observable phenomena. On the other hand, complementarity does not require us to treat models merely as useful fictions, or to accept a positivist interpretation. Complementarity when understood in this way is not inconsistent with critical realism.

3. Numinous and Mystical Experience

Within physics complementary models are used in the domain of the unobservably small, whose characteristics seem to be radically unlike those of everyday objects; the electron cannot be adequately visualized or consistently described by familiar analogies. Perhaps within religion also there are inherent limitations in the applicability of visualizable models and familiar analogies. We shall see whether personal and impersonal models of Ultimate Reality may be thought of as complementary representations. In accordance with the principle that we should start from the experiential basis of religion, we must consider first the fundamental types of religious experience which give rise to these two broad types of models.

There are two common types of religious experience which Ninian Smart has traced among world religions: numinous encounter (associated with worship) and mystical union (associated with meditation).13 Numinous encounter received its classic description in Rudolf Otto’s The Idea of the Holy. Its characteristics include a sense of awe and reverence, mystery and wonder, holiness and sacredness, fascination and dread. Typical examples are the vision of Isaiah in the temple, the call of Paul or Muhammad, or the theophany of Krishna in which Arjuna is struck dumb in amazement. Numinous experience is often accompanied by moral demand and a response of humility. It is institutionalized as worship, and as ritual ascribing value to the object of worship and expressing the inferiority of the worshipper (often symbolized by bowing down).

Smart asserts that numinous experience is usually interpreted by personal models of God. The worshipper thinks of God as distinct and separate from himself. The overwhelming character of the experience suggests an exalted view of the divine. The ‘distance’ between man and God, acknowledged in self-abasement, leads to an emphasis on the ‘otherness’ and transcendence of God (rather than to divine immanence as in the mystical tradition). The sense of being grasped and laid hold of, the unexpectedness and confrontation, and the conviction that one’s response is evoked, all seem to point to an activity independent of man’s own control, a divine rather than human initiative. Winston King, after describing the ritual-communal component of religion which expresses ‘the gap between worshipping man and the worshipped Ultimate’, concludes: ‘We have here a root of personalistic theism, whatever may be the doctrine of the given tradition. For the ritualized forms of relationship tend to be those of personalized worship: sacrifice and petitionary prayer to a deity who will hear and heed.’14

The second type of experience, mystical union, also seems to have common characteristics, or at least ‘family resemblances’, in cultures which had almost no historical interaction. W. T. Stace has documented such recurrent traits as unitary consciousness and intense joy.15 In some cases the unity is found in the world, the ‘one’ in the multiplicity of objects, as in nature mysticism. In other cases, the world is left behind on the inward route of contemplative discipline; undifferentiated unity is found in the depth of one’s own soul. Despite such variations, there is considerable agreement in these descriptions. The Christian mystic can recognize the Hindu mystic’s experience as similar to his own. Here the typical expression is meditation and contemplation rather than worship or ritual. Discipline, and sometimes asceticism, leads to peace, serenity and bliss. Sacrament and scripture are left behind in enlightenment, immediacy of knowledge, realization of unity, and liberation from the illusion of separation. In identity with the One beyond time and space, all differences are obliterated. All dichotomies (human-divine, subject-object, time-eternity, evil-good) are transcended.

The mystic is cautious in the use of models; he is likely to stress the ineffability of the experience. He may start from the via negativa, the assertion of what the divine is not; Brahman, says the Hindu, is neti, neti (not this, not this). But he usually does not stop there. ‘The mystic must be silent: but he cannot be: he must speak.’16 The images and models he does use are sometimes personal. He may speak of ultimate reality as a Self identical in essence with the individual self; or as the World Soul with which one’s own soul is merged. Atman, man’s true self, is Brahman, the divine found within; ‘That art thou.’ But mystical experience has more often been associated with impersonal images. The self is absorbed in the pantheistic All, the impersonal Absolute, the divine Ground. The distinction between subject and object is overcome in an all-embracing unity beyond all personal forms. The self loses its individuality ‘as a raindrop loses its separate identity in the ocean’.17

Ninian Smart has shown that although Western religious traditions have been predominantly numinous and Eastern traditions predominantly mystical, all the major world religions have in fact included both types of experience.18 Early Israel gave priority to the numinous; biblical literature portrays the overwhelming sense of encounter, the prophetic experience of the holy as personal, the acknowledgment of the gulf between the worshipper and the object of worship. But in later Judaism and Christianity there are many typically mystical writings. Islam in its early history was also strongly numinous, but later developed its own versions of mysticism, especially in Sufi literature. On the other hand, early Buddhism was predominantly mystical, following the meditative path, but Mahayana Buddhism includes the numinous strand in the worshipm of the heavenly Buddhas and Bodhisatvas. Evidence of both strands is present in all the major religious traditions. We shall explore them as a possible case of complementarity.

Although religious experience as described by Rudolf Otto is predominantly numinous, it includes a component which when more fully developed seems to lead towards mysticism. Otto spoke of the mysterium tremendum et fascinans, the mystery which evokes fear and awe, and yet also attracts and fascinates. Otto himself correlated this polarity with the tension between divine wrath and divine love, which is undoubtedly a characteristic form of interpretation in the West. But a more universal typology might correlate the tremendum pole with the prophet’s feeling of the unapproachability of God, the radical discontinuity of Creator and creation, and the response of humility and prostration; man must keep his proper distance, of which any violation is tabu. The fascinans pole, on the other hand, resembles the mystic’s feeling of the nearness of the divine, the continuity of all things, the participation of man in the ultimate identity. The polarity of withdrawal and approach, or distance and identity, seems to be present within the experience of the sacred, though for different individuals and traditions one aspect or the other may be more prominent. After discussing the prophetic and mystic traditions rooted in these two aspects of the experience of the sacred, Conrad Hyers concludes:

If pressed to their logical conclusions on the basis of their respective premisses they become mutually exclusive visions which can only meet in tragic contradiction. Nevertheless, inasmuch as each articulated one side of a basic religious polarity, they are necessarily complementary visions the solution to such historical oppositions and antagonisms is, therefore, a dialectical one -- not in the Hegelian or Marxian sense of dialectic, but through a dialectic which acknowledges both sides of those paradoxes intrinsic to the religious situation. Such a dialectic is not strictly a synthesis, and certainly not an eclectic juxtaposition of elements; it is the recognition and realization of the implications of a fundamental duality in the presentation of and response to the sacred.19

Winston King has discussed the universal polarity of personal and impersonal symbols:

By and large the Eastern religions can be distinguished from Western religions specifically by their emphasis upon the impersonality of the ultimate Object of religious devotion. The terms used for ultimate realities in the East -- Erahman, Nirvana, Tao, Heaven, and Kami -- are almost entirely non-personal in connotation and experience. In the East, the personal form of existence and individual consciousness are looked upon as inferior and limited; they exist at a level of being and experience that does not accurately describe the truly Ultimate Reality.20

But King also points out that personal symbols of relation and impersonal symbols ofidentilication co-exist in all the major traditions:

We will see how mingled the relational-personal and the identificatoryimpersonal quality of symbols actually is in both East and West religious traditions. In anticipation: the non-God of the East often achieves personalistic forms and is so worshipped; and the God of the West has been more than once conceived of as impersonal reality.21

As an example, we shall consider the presence of both types of religious experience, and the corresponding interpretive models, in Hinduism. The numinous type recurs throughout the Bhagavadgita. The bhakti tradition has followed the path of devotion, the way to God through worship, loving adoration, and reliance on divine grace. To his devotees, Vishnu was the one God, the Supreme Person beyond the imperishable Brahman. For Ramanuja, the personal God was the Supreme Lord, worthy of dedicated worship, of whom the impersonal is only one manifestation. This approach is predominant in daily life and practice in India today. But the mystical approach is also prominent in the history of Hinduism. In Sankara’s monism, there is one supreme Brahman, the Absolute without attributes. In the higher state of liberation through meditation, there is no ‘other’ to worship; the representation of God as Creator (Isvara) is merely a useful way of portraying the divine for the ordinary worshipper. Similarly, Vedanta Hinduism speaks of two levels of truth: the highest truth of One Ultimate Reality, known in contemplation, and a lower level in which Brahman can be viewed as personal creator and focus of worship.

A modern interpreter, S. Radhakrishnan, comments on the tolerance of Hinduism in allowing both personal and impersonal representations of the divine. They are, he says, two sides of the same reality, not incompatible claims or merely different human viewpoints:

Western forms of religion are inclined to hold that one definition is final and absolute and others are false. In India, each definition represents a darsana or viewpoint. There are many different ways of viewing one experience. The different darsanas are different viewpoints which are not necessarily incompatible. They are pointers on the way to spiritual realization.22

Radhakrishnan holds that we can apply personal terms since the personal is the highest category we know in finite experience. Yet personal terms ‘do not tell us about God in himself but only what he is to us’, but also as he ‘expresses himself in a personal mode’. It appears that Radhakrishnan’s view is finally monistic; the mystic’s awareness of his spiritual identity with the suprapersonal Absolute is a higher goal than the worship of a personal God.

Vedanta Hinduism allows a place for theism, but tends to see it as a lower stage of spiritual development. The mystical strand is given priority over the worship of the personal, holy divinity. At the highest level of truth, the personal in both man and God is swallowed up in the impersonal Absolute. Moreover, the impersonality of the cosmic order in Hindu thought is reinforced by the idea of karma as an impersonal moral law concerning the inexorable consequences of one’s deeds. From its side, Christianity has made room for mysticism but has preserved the dominance of the theistic framework. It has been critical of the extremes of pantheistic interpretation. It has found unacceptable any total identity which obliterates the gap between man and God, as in Meister Eckhart’s statement: ‘If I am to know God directly, I must become completely He and He I, so that this He and this I become and are one I.’23 As analogy for the union of man and God, Christian mystics invoke the interpersonal unity of marriage more often than the impersonal merging of a drop in the ocean. Worship, rather than meditation, remains the basic Christian experience.

This relative priority of personal or impersonal models has far-reaching implications. Only with a personal God can there be divine initiative and freedom rather than cosmic necessity. Divine initiative, together with the ontological and epistemological distance assumed between man and God, is a correlate of the ideas of historical revelation, grace and redemption; the gulf can only be bridged from the side of the divine. Belief in a purposeful creator active in history has led to the conviction that the historical process is directional rather than cyclical. Again, Western traditions have given greater stress to human individuality (which in extreme forms becomes an anti-social individualism). Belief in the worth of the individual is in part based on an understanding of the value of persons to God. Man’s freedom can of course be jeopardized as much by an omnipotent God as by a pantheistic determinism, but the West has usually viewed the self as an active agent. It has found a larger place for ethical activism, whereas the oriental quest for inner peace has more often led to quietism -- though one can only admire the compassion and sensitivity of many of its saints and holy men. These relationships are of course very complex. They are mentioned here only as a reminder that any model functions in a total network of ideas and attitudes which form an interrelated and organic whole.

These differences between integral religious traditions -- Hinduism and Christianity, for example -- are so great that they can best be understood as the product of different paradigms. In the following three chapters I will analyse a paradigm as ‘a tradition embodied in historical exemplars’ and show how it dominates the patterns of life and thought of a scientific or a religious community. Thus I would propose that we should not refer to the Hindu Brahman and the Christian God as complementary models (if some analogy with quantum physics is thereby implied), since they are not used in the same paradigm community. However, the use of personal and impersonal models within the Hindu tradition, or within the Christian tradition, does seem to present some interesting parallels with complementarity in physics, which we must now examine further.

4. Personal and Impersonal Models

Consider first the suggestion that two different personal models within a given tradition might be thought of as complementary. William Austin has developed Bohr’s proposal mentioned earlier that the idea of complementarity may be applied to divine love and divine justice in biblical thought. After describing in some detail the principle of complementarity in physics, Austin suggests that images of God as Father and as Judge are complementary models used to interpret individual and corporate experience.24 The prophet Amos, he points out, interpreted events in Israel’s history primarily in terms of God’s judgment, while Hosea understood events in terms of God’s forgiveness. One possible explanation would be that these different models were used in different historical circumstances, just as there are ‘different experimental situations’ in physics. But it appears that they actually faced very similar historical situations, and both prophets made some use of both models. Even for Amos there was hope of forgiveness, at least for a remnant; and for Hosea, Israel was judged by the very love which would not abandon her. Here is a kind of mutual limitation, Austin proposes, in which the presence of one model prevents the exclusive development of the other. The models are taken in interaction rather than in total independence, and there is a tension between them as there is in quantum physics.

I suggest, however, that in this case a compromise model can be introduced. In human life, a loving father must remember the demands of justice, or else concern for his child becomes sentimentality. And a judge must have some scope for mercy, or else justice will become legalistic retribution. So the models of God as Father and as Judge can to some extent be merged, or perhaps included in a wider image -- an ideal King, for example, who is both loving and just to his subjects. Moreover the guiding theme of the covenant provides a historical framework within which both love and justice have a place, even though in any particular situation their demands appear to be opposed. I do not see here quite the kind of mutual exclusiveness that exists between particles and waves, which prevents the development of a single compromise model in quantum physics.

A case which seems more analogous to complementarity in physics is Paul Tillich’s use of personal and impersonal symbols, to which the remainder of this chapter is devoted. He acknowledges the importance of personalistic imagery in the biblical tradition. The Holy is known in the divine-human encounter, the person-to-person relation of mutual freedom, reciprocity and individuality. In guilt, forgiveness and faith, man responds to a God understood as a separate being who acts to save. But Tillich also frequently uses the impersonal symbols associated with mysticism. He presents ‘individuation and participation’ as one of the basic polarities which characterize all existence. Whereas I-Thou encounter symbolizes the individuation of man and God, the mystic’s experience of unity points to participation in the divine. ‘Individuation’, taken alone, makes God and man separate beings under the sway of the subject-object distinction, but ‘participation’ alone leads to the loss of self-hood and freedom.25

Tillich finds evidence of the personal-impersonal polarity throughout biblical thought. If the Word of God is personal address to man, it is also the Logos, the rational structure of the cosmos, the universal creative power. If revelation is God manifest in unique historical events, it is also the divine self-manifestation in the depth of all events. If prayer is man’s personal address to God, it is also surrender to that which is working in us, nearer than we are to ourselves. If love is reconciliation overcoming estrangement between personal beings, it is also the reunion of that which was separated, the recovery of a fundamental identity.26

Theistic supernaturalism, says Tillich, pictures God as a being, separate from all other beings. It portrays him by merely extending the categories of finitude -- spatially, as above the world; temporally, as before the world; and causally, as a cause among other causes. But pantheistic naturalism makes God only a power within the world, ignoring ‘the decisive element in the experience of the holy, namely the distance between finite man, on the one hand, and the holy in its numerous manifestations, on the other’. Pantheism also neglects the mutual freedom of man and God, the freedom of the created ‘to turn away from its essential unity with its creative ground’.27

For Tillich, God is not a being, but being-itself Some critics have assumed that ‘being’ is a lifeless, static, impersonal concept. But Tillich stresses the active power of being overcoming non-being, the power of life resisting the threat of death. God is the creative ground of being, the transcendent source of vitality and dynamics as well as of form and structure. God is not a person, for to Tillich an individual personal centre or self implies a radical separateness from everything else. But God is the ground of self-transcendence and personhood which is actualized in finite persons:

‘Personal God’ does not mean that God is a person. It means that God is the ground of everything personal and that he carries within himself theontological power of personality.... Ordinary theism has made God a heavenly, completely perfect person who resides above the world and mankind. The protest of atheism against such a highest person is correct. There isno evidence for his existence, nor is he a matter of ultimate concern. God is not God without universal participation.28

Elsewhere Tillich writes:

This means that being and person are not contradictory concepts. Being includes personal being; it does not deny it. The ground of being is the ground of personal being, not its negation.... Religiously speaking, this means that our encounter with the God who is a person includes the encounter with the God who is the ground of everything personal and as such not a person. Religious experience, particularly is expressed in the great religions, exhibits a deep feeling for the tension between the personal and the nonpersonal element in the encounter between God and man. The Old as well as the New Testament has the astonishing power to speak of the divine in such a way that the I-thou character of the relation never darkens the transpersonal power and mystery of the divine, and vice versa.29

In the basic polarities of existence -- individuation and participation, dynamics and form, freedom and destiny -- the first term of each pair is more personal’. But both terms in each polarity must be included in the characterization of the divine life.

Any aspect of being can be a symbol of God since all being participates in being-itself. Religious symbols, says Tillich, participate in the reality they represent and thus point beyond themselves. They combine the concreteness of the specific events or objects from which they are drawn with the ultimacy of the ground which they symbolize. They are affirmed as symbols but denied as literal predications, for the categories of finitude cannot be applied literally to God. Use of a religious symbol demands self-criticism and awareness of its limitations. Nothing finite deserves worship; to elevate the symbol itself to ultimacy is idolatry. Literal language is also unable to express the existential involvement which is a precondition of apprehension of the divine. The use of a symbol, personal or impersonal, requires a dialectic of affirmation and negation.

Tillich suggests several criteria for evaluating religious symbols, in addition to this capacity for self-negation.30 A symbol of the ultimate must transcend the subject-object division, for the characteristics of being-itself are equally present in human life and beyond it; the symbol must express the basic unity of all things, of which man is aware in the depths of his own being. Further, a symbol can be judged by its integrating or disintegrating power in practice. A symbol of the true ultimate will unify men as individuals and as communities rather than dividing them. Its effects will be creative rather than destructive and it will vindicate its promises of human fulfilment. A final criterion is the adequacy of the symbol to religious experience -- its authenticity in expressing the state of a person grasped by an ultimate concern and in representing that which concerns him ultimately. Thus Tillich’s theory of symbols stresses their religious use in expressing and evoking experience, and their role as vehicles of commitment and devotion.

Do personal or impersonal symbols predominate in Tillich theology? To answer this, we must distinguish three sources of his theology whose influence varies among his diverse writings. First, there is the experiential basis of his thought. Here there is much in common with the impersonal features of the mystical tradition. In the heritage of the Neo-Platonic mystics, Jacob Boehme, and German mysticism, Tillich frequently talks about immediate awareness, the union of knower and known, the intuition of identity, and participation in a unity beyond the subject-object division. Man is grasped by the holy, the unconditional, the sacred. Tillich defines faith not as person-to-person trust but as ‘ultimate concern’, unconditional demand, the final ground of a person’s values and the justification for his decisions. Whatever its object, ultimate concern is a total perspective involving the whole person and requiring his unreserved allegiance. Tillich holds that theology must start from the ‘questions’ implied in human existence and the ‘answers’ experienced in human life in response to revelatory events. One such experience is courage overcoming the anxiety of finitude, temporality and meaninglessness. In accepting ourselves as accepted we participate in the self-affirmation of being-itself; we seek the sustaining power which undergirds and supports our courage to be in the face of the threat of non-being. Another basic experience is reconciliation overcoming separation and estrangement. The redemptive power of love is known in human life. Grace and redemption are not theological abstractions but experienced realities in which divisions within man, and between man and his neighbour, are healed. Salvation is literally ‘being made whole’ by the healing forces at work in the world. God is the structure of reality and the power of being which brings about these transformations of human existence, which can be described in personal terms as response to love and forgiveness and in ontological terms as the reunion of the separated.

In the second source of Tillich’s theology, the biblical tradition, personalistic symbols are more strongly represented. His description of reconciliation is of course indebted to biblical formulations, but his presentation of Christ as the New Being in whom is manifest the power of love brings out more fully the importance of personal symbols. Tillich acknowledges that the divine is manifest to particular communities in the concreteness of revelatory events and persons. For the Christian community, the cross is the supreme symbol, for in his self-sacrifice Christ pointed beyond himself and surrendered the particular to the ultimate; the cross was the manifestation of God’s participation in man’s existence, universally present but not universally recognized.31 Tillich’s own background in the Lutheran Church and his sensitivity to Luther’s experience of guilt, forgiveness, personal faith and divine grace, are reflected at many points in his writings, especially in his sermons.32

But the third source of Tillich’s thought, the tradition of German idealist philosophy, gives greater emphasis to impersonal conceptions and is more evident in his systematic discussion of the doctrine of God.33 He does not accept Hegel’s vision of participation in the all-inclusive Absolute in which all differences are overcome in harmonious synthesis. Rather, with Schelling, he conceives of the finite as both participating in and estranged from the infinite. Revelation must come to man, and yet it comes not as something alien and foreign. Estrangement is fragmentarily overcome by love, not totally resolved by rational synthesis. Yet for Tillich, man’s essential unity with the infinite is never destroyed by man s actual estrangement. The basic identity of thought and king, the unity of subject and object, and the possibility of immediate awareness by participation, which are assumptions of Western idealism from Plato to Hegel, are all fundamental to Tillich’s viewpoint. The ontological structures of the world, the unity underlying polarity, are the clues to understanding God. The differentiation of separate selves is not a reflection of the goodness of creation but a source of ambiguity and division. When man discovers God he finds something identical with himself, not a stranger or an inference at the end of an argument.34

It seems to me, then, that even though Tillich’s theory of symbols dwells on religious uses, and his theological method starts from existential questions, his formal discussion of God is more strongly indebted to idealist philosophy than to either religious experience or the biblical tradition. In his other writings, models of God serve mainly religious functions, but in his systematic doctrine of God models serve mainly metaphysical functions and are greatly influenced by a philosophical ontology in which impersonal categories predominate. God may be transpersonal (beyond the distinction between personal and impersonal), but our models must use analogies from the life with which we are familiar, and for Tillich it is the impersonal structures which are emphasized. In Chapter 8 below I will suggest that process metaphysics is more congenial to the personalistic models characteristic of biblical thought, while including impersonal features of the cosmic evolutionary process.

In concluding this chapter we must ask whether personal and impersonal models in such a theology as Tillich’s could be considered complementary. I can see a certain parallel with the situation in atomic physics: the use of two models which cannot be combined, along with recognition of the limitations of all models and the inadequacy of literalism. Is there a greater contrast between personal and impersonal models than between waves and particles? Perhaps there are fewer common properties in the first polarity than in the second, but we should not minimize the divergence in conceptualities between waves and particles. (One might even speculate that particles represent ‘individuation’ in separate units, whereas waves represent a kind of ‘participation’ in a more inclusive field.) The main parallel is simply that in both situations two contrasting types of models are used.

There are also features of complementarity in atomic physics which are absent from theology. In atomic physics there is a unifying mathematical formalism which allows at least probabilistic prediction of particular observations. There is consistency at the level of theory, though not at the level of models. Theory specifies what is essential in the models by indicating the positive and negative analogies. In theology, doctrinal schemes provide some conceptual unity in the quest for coherence, and they serve a function not unlike that of theories. But their relation to experience is more ambiguous, and no one would claim for them any kind of predictive power on even a probabilistic basis. We will explore theory and observation in science and religion in the next two chapters, before returning to models in Chapter 8.


1. See, for example, Philipp Frank, Philosophy of Science, Bailey Bros & Swinfen and Prentice Hall 1957, chap.9. A very readable account is given in Richard Feynman, The Character of Physical Law, MIT Press 1967.

2. Ernest Nagel, The Structure of Science, pp. 300, 302.

3. Mary B. Hesse, Forces and Fields, Thomas Nelson and Sons 1961, p.23. See also Models and Analogies in Science, pp. 57ff.

4. Ernest H. Hutten, The Language of Modern Physics, Allen & Unwin and Macmillan 1956, p.164.

5. Mary Hesse in D. Bohm et al., Quanta and Reality, World Publishing Co. 1964, p. 57.

6. Niels Bohr, Atomic Theory and the Description of Nature, Cambridge University Press 1934, p.96.

7. Niels Bohr, Atomic Physics and Human Knowledge, Chapman & Hall and John Wiley & Sons 1958, pp. 39ff.

8. Ibid., pp. 92ff. See also Gerald Holton, ‘The Roots of Complementarity’, Daedalus, vol.90, 1970, p.1015; J. Robert Oppenheimer, Science and the Common Understanding, Oxford University Press and Simon and Schuster 1954, chaps. 4-6.

9. C. A. Coulson, Science and Christian Belief, Oxford University Press and University of North Carolina Press 1955, chap. 3.

10. D. M. MacKay, ‘Compleinentarity’, Aristotelian Society Supplementary Volume 32, 1958, p.105..

11. P. K. Feyerabend, ‘Problems of Microphysics’, in R. G. Colodny (ed.), Frontiers of Science and Philosophy, Allen & Unwin and University of Pittsburgh Press 1962.

12. Peter Alexander, ‘Complementary Descriptions’, Mind, vol. 65, 1956, p.145.

13. Ninian Smart, Reasons and Faiths, Routledge & Kegan Paul 1958; World Religions: A Dialogue, Penguin Books 1969; The Concept of Worship, Macmillan 1972; ‘Revelation, Reason and Religions’ in Ian Ramsey (ed.), Prospect for Metaphysics, Allen & Unwin 1961.

14. Winston King, Introduction to Religion: A Phenomenological Approach, Harper & Row 1968, p. 165.

15. W. T. Stace, Mysticism and Philosophy, Lippincott 1960; see also William James, Varieties of Religious Experience, Collier-Macmillan and Random House n.d.

16. G. van der Leeuw, Religion in Essence and Manifestation, p.501.

17. Versions from Omar Khayyam, Mahmud, and the Upanishads are quoted in ibid, p. 498.

18. See especially Reasons and Faiths.

19. M. Conrad Hyers, ‘Prophet and Mystic: Toward a Phenomenological Foundation for a World Ecumenicity’, Cross Currents, Fall 1970, p.435.

20. King, Introduction to Religion, pp.20-2l.

21. Ibid., p.152.

22. 5. Radhakrishnan, Recovery of Faith, AlIen & Unwin 1956, p. 155

23. Eckhart, Mystische Schriften, p.122, quoted in Evelyn Underhill, Mysticism, Methuen & Co. 1911, p.418.

24. William Austin, ‘Waves, Particles and Paradoxes’, Rice University Studies, vol. 53, 1967, pp. 93ff.

25. Paul Tillich, The Courage to Be, Yale University Press 1952, chap. 6.

26. Paul Tillich, Biblical Religion and the Search for Ultimate Reality, University of Chicago Press 1955.

27. Paul Tillich, Systematic Theology, vol. II, University of Chicago Press 1957, pp.5-10.

28. Paul Tillich, Systematic Theology, vol. I, University of Chicago Press 1951, p.245.

29. Biblical Religion and the Search for Ultimate Reality, pp. 83-84.

30. Paul Tillich, Dynamics of Faith, Harper & Row 1957, chap. 3; also ‘The Meaning and Justification of Religious Symbols’, in Sidney Hook (ed.), Religious Experience and Truth, New York University Press 1961.

31. Systematic Theology, vol. II, pp. 97-180.

32. Paul Tillich, The Shaking of the Foundations and The New Being, Charles Scribner’s Sons 1955.

33. Systematic Theology, vol. I, Part 11.

34. Paul Tillich, ‘Two Types of Philosophy of Religion’, in Theology of Culture, Oxford University Press 1964.